signal amplitude: dry Ultracortex with P300 or cVEP [resolved]

jetjet U.S
edited July 26 in Electrodes

Hello there,

I bought the Ultra "Mark IV" EEG Headset around 3 months ago. Ever since the last time I posted on this forum and even before that, the quality of the signals that come from my headset is not good.

The issue is that even when using OpenVibe to classify the signals as P300 or not, the classifier does not detect any P300 signals. The potential issues that came to mind were latency/timing issues, how worn out the electrodes are, and the signal quality that come from the electrodes themselves.

The impedance values according to other forums, were not bad on the OpenBCI GUI.

From the last post I put on this forum, I was told to change the latency settings in the Ubuntu 16.04 VM I had, which I did; however, the classifier from OpenVibe still did not predict any P300 signals. Therefore, I chose the suggested alternative MindAffect software. I relocated the electrodes to the back of the headset, as the directions stated.

After running the SNR test from their Python API on my base Windows system, the values were always too high, unless a drastic change was made to the headset at the moment-- for example, moving the headset rapidly to the side. I wondered if this was an issue with being too close to other devices that may have interrupted the signals, so I moved to a different room and ran the test; however, the SNR stayed the same. The SNR is always high, regardless of whether the headset is on or off. I also wonder if the issue is that I am utilizing Dry Comb electrodes, instead of wet electrodes for this task.

It is hard to pinpoint what the exact issue is, which is why I came to this forum.

Comments

  • wjcroftwjcroft Mount Shasta, CA
    edited July 1

    Jet, hi.

    Past experience with other P300 users, shows that wet electrodes (paste, gel+cap, or saline / water pads) are preferred. This is because of the small signal amplitude of the P300 signal. It is SO small, that it requires signal averaging multiple epochs, to increase the signal to noise level. Past Cyton users have had success with wet electrodes, such as,

    http://blog.jfrey.info/

    And as far as MindAffect goes, they too have best success with their own headband, which is saline pad (or even water + pad) based. So very easy to use and no sticky stuff or cleanup required. MindAffect VEP signal at the occipital cortex is larger, but because of the very high modulation rate of the visual stimuli, good accuracy is required of BOTH the EEG, AND the display generation hardware. Did you do their test with the photoresistor to verify your display setup?

    https://github.com/mindaffect/pymindaffectBCI [FrameRate check}
    https://mindaffect-bci.readthedocs.io/en/latest/build_your_own_optical_sensor.html
    https://mindaffect-bci.readthedocs.io/en/latest/stimulus_time-lock_test_tutorial.html

    Regard, William

  • jetjet U.S

    Hello there,
    I have been using the dry comb electrodes for my UltraCortex Mark IV headset. However, I have been receiving poor results with MindAffect; I have not been able to free type for predictive letter typing. Only on one occurrence was I able to type out the letter h, e, l, and o, but immediately afterwards the program was not able to type out any letters at all.
    I have done the frame rate check, and it says that the delay is 7.0+-0.3 [5, 19]; therefore, I suspect that the issue is not due to timing, but due to the contact of the electrodes. I have checked the impedance values, and even with 1k kOhm as the max impedance value of all the electrodes, the program still was not able to predict any letters.
    This brings me to my main point: is there a wet electrode that is compatible with the UltraCortex Mark IV headset? and if so, what is it?
    This is because I found that the documentation does not show that there is a wet electrode that is compatible with my headset.

    Thank you.

  • wjcroftwjcroft Mount Shasta, CA

    @jet, hi.

    I merged your new thread into this existing thread on the same subject.

    As I suggested previously, some applications such as P300 or cVEP, require more signal amplitude than dry comb electrodes can provide. Have you tried using a wet electrode approach, either cups and paste, an electrode cap, or best I believe with MindAffect, their own printable water or saline based wet headband? It's even possible if you contact them they might have pre-built bands available.

    I don't see any straightforward way to integrate moisture with the dry electrodes used with Ultracortex. And moisture is what increases the signal amplitude with the skin / electrode interface. Impedance is not the issue, it's that dry electrodes (especially passive dry as opposed to active dry), just do not provide enough signal amplitude.

    Recent post by Sapien Labs,

    https://sapienlabs.org/eeg-signal-quality-in-wet-versus-dry-electrodes/

    Regards, William

Sign In or Register to comment.